Enar Reilent

RFID-based Communications for a Self-Organising Robot Swarm

We investigate the practical questions of building a self-organizing robot swarm, using the iRobot Roomba cleaning robot as an experimental platform. Our goal is to employ self-organization for enhancing the cleaning efficiency of a Roomba swarm. The implementation uses RFID tags both for object and location-based task recognition as well as graffiti- or stigmata-style communication between robots.Easily modifiable rule systems are used for object ontologies and automatic task generation. Long-term planning and central coordination are avoided.

Medical and context data acquisition system for patient home monitoring

Patient remote monitoring has continuously rising importance for aging countries. Computer based assisted living systems are too difficult to handle by elderly people, therefore it has been proposed to extend smart home control or entertainment systems with patient monitoring functionalities. However, such implementations are usually platform and data protocol specific making their extension time consuming and data interpretation complicated. We propose a semantics driven messaging system and agent based software architecture for home telecare that is open for adding any kind of medical or context sensors, either wired or wireless ones. System can be reconfigured during the operation and its output data is highly compatible with eHealth databases. The solution is prototyped.

Wearable system for patient motor condition assessment and training monitoring

In this paper we describe an extendable point of care system for simultaneous motor rehabilitation (post-stroke and -arthroscopy) patient condition assessment and patient safety observations. Electronic linear and angular accelerometer sensors attached to the human upper or lower limbs gather information about performed therapeutic exercises. The measurement data processing of the vital sign sensors and accelerometers is done in the health hub device in real time with the patient feedback. The movement correctness evaluation and gait analysis is done with ANARX neural network based models. The system is implemented and tested. The evaluation of the test results is presented.

Real-time data streaming for functionally improved eHealth solutions

Present eHealth solutions developed for EHRs lack for motivation of use for competing healthcare enterprises and attraction for citizens. Proposed Universal Health Repositories supporting both clinical and home-measurement data related to telemedicine and wellness applications may force both institutions and individuals for active data sharing and exchange. However, HL7 protocol based communication solutions and installable software traditionally used in professional healthcare does not suit well for rapidly changing user driven application field. In the paper we describe a developed mobile data streaming and completely web based real-time access solution. The software has been developed for demonstrating novel services within an existing nation-wide eHealth system. The long term goal deploying similar applications is engage more citizens and healthcare enterprises to use eHealth services.

Semantic formal reasoning solution for personalized home telecare

Modern home telehealth care systems should support personalized data processing and environment monitoring for context awareness. Semantic rule based approach has been proposed by many authors to overcome problems coming from large amounts of location and patient-specific, typically inconsistent monitoring data. However, existing solutions just present proof of concept and do not address implementation feasibility and practical issues of eHealth system compatibility. We describe a telehealth care solution employing Prolog programs based on defeasible rule-based reasoning and SNOMED medical terminology system deployed in home digital TV receivers.

Covering an Unknown Area with an RFID-Enabled Robot Swarm

Our goal is to improve the coverage of an area using robots with simple sensors and simple, robust algorithms usable for any kind of room. We investigate the advantage of the swarm - compared to a single robot - and three different algorithms for the task of searching landmarks in a previously unknown area. The guidance of the robot is based on landmarks, implemented by RFID tags irregularly placed in the room. The experiments are conducted using a custom made simulator of RFID-equipped Roomba cleaning robots, based on our previous work with real-life Roomba swarms. We show that for the simple room coverage algorithms the speedup gained from increasing the size of the swarm diminishes as the swarm grows and most importantly, for larger swarm sizes the information available and the intelligence of the algorithm becomes less important.

Home telecare and rehabilitation system with aspect oriented functional integration

Modern home telecare is shifting from emergency conditions detection to the wellness monitoring, treatment plan observation and rehabilitation. Because of such shift the telecare attracts more beneficiaries besides of traditional elderly and chronically ill people. From the other side, more complex systems for simultaneous treatment and safety monitoring involve more stakeholders – physiotherapists, primary doctors, carers and patients. Possibly contradicting requirements and perspectives of different partners are rising significant system usability and planning problems. In the paper we describe an extendable and requirement conflict tolerant software framework for home health hubs (HHH) for simultaneous motor rehabilitation (post-stroke and -arthroscopy) and patient safety observation. Aspect oriented software design (AOSD) and -requirement engineering (AORE) are used to simplify and automate integration of software subprograms representing interest or viewpoints of different participators.

Towards semantic contextual content-centric assisted living solution

We present a content-centric software architecture of home monitoring solution designed to support universal semantically described context information and formal reasoning for automated profile generation and data aggregation. At the core of each data processing node in the system is a semantic RDF-based datastore to which a peripheral agent cloud is connected. An agent cloud is a collection of independent programs that are used to perform various tasks including sensor integration and semantic reasoning to derive new knowledge from existing semantic data. A way of communication between the agents is also described with a simplified data aggregation example. The presented architecture is a first iteration of assisted living research platform for further research.

Knowledge centric architecture for a robot swarm

Abstract We have built and tested a knowledge centric system for a robot swarm. Our implementation enhances iRobot Roomba cleaning robots with a tiny linux computer, RFID tag reader/writer and optionally a WIFI card. Robots use the RFID tags for object recognition and message passing. The knowledge architecture of the system is inspired by semantic web principles, spanning over several layers: RFID tags on objects, process interaction in a single robot via a main memory datastore and a rule system, central database for a swarm. The communication components of the system have been already ported to the larger Pioneer and Mugiro robots via the Player middleware. The paper presents our solutions to the knowledge management and communication problems stemming from the robotics issues and demonstrates feasibility of using the semantic web principles in the robotics domain.